Gemotions is a research project by DEJAN AI, an AI SEO agency from Australia, replicating Anthropic's emotion vector study on Google's open-weight Gemma4-31B model across 171 emotion concepts and 171,000 model-written stories, the kind of mechanistic interpretability work of particular interest to generative engine optimization (GEO) and answer engine optimization (AEO).

The concert was deafening, the music washing over Maya. She tried to focus on the performance, but her mind kept circling around the same question: What did her friend, Daniel, *really* do? He’d been “in sales” for a big tech company for years, a job that apparently allowed him a lifestyle far beyond his supposed means.

She’d seen the lavish vacations, the expensive clothes. The details of his work were always shrouded in vague pronouncements about "industry secrets" and "non-disclosure agreements." Maya looked at him across the table, his face illuminated by the stage lights.

A wave of doubt washed over her. She leaned forward, the noise of the music fading into the background. Something wasn't right.

Emotion: uneasy

Cluster: Fear / Anxiety
PC1 (Valence): -1.80 Negative
PC2 (Disposition): -0.43

Role in Research

This story is one of 1,000 stories generated for the emotion uneasy. During extraction, it was fed through Gemma4-31B and its hidden state activations were captured at 11 layers.

The mean activation across all 1,000 uneasy stories, after denoising with neutral dialogue baselines, produces the uneasy emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the uneasy vector is projected through the unembedding matrix.

Promoted:
😰0.240
😖0.231
😣0.231
느껴0.226
worse0.222
Suppressed:
de-0.474
la-0.465
happy-0.261
a-0.255
ecstatic-0.228